(SYB) 12-1

Tracking and Control of Gas Turbine EngineComponent Damage/Life

Dr. Link C. Jaw, Dr. Dong. N. Wu, Dr. David J. BrygScientific Monitoring, Inc., 4801 S. Lakeshore Dr., Tempe, Arizona 85282, U. S. A.

link@scientificmonitoring.com, http://www.scientificmonitoring.com

ABSTRACT

This paper describes damage mechanisms and the methods of controlling damages to extend theon-wing life of critical gas turbine engine components. Particularly, two types of damagemechanisms are discussed: creep/rupture and thermo-mechanical fatigue. To control thesedamages and extend the life of engine hot-section components, we have investigated twomethodologies to be implemented as additional control logic for the on-board electronic controlunit. This new logic, the life-extending control (LEC), interacts with the engine control andmonitoring unit and modifies the fuel flow to reduce component damages in a flight mission.The LEC methodologies were demonstrated in a real-time, hardware-in-the-loop simulation. Theresults show that LEC is not only a new paradigm for engine control design, but also a promisingtechnology for extending the service life of engine components, hence reducing the life cyclecost of the engine.

1. Introduction

Gas turbine engines consist of primarily rotating components. These rotating componentsoperate under cyclic loading condition and harsh environment (i.e., under high temperature,pressure, corrosion condition) such that the deterioration of these components is accelerated.Deterioration is generally tracked by damages, or damage rates, for different damagemechanisms. The most common damage mechanisms for a gas turbine engine include: low cyclefatigue (LCF), thermo-mechanical fatigue (TMF), high cycle fatigue (HCF), creep, rupture,corrosion, and foreign object-induced damages (FOD). Of these common damage mechanisms,LCF and HCF are primarily design issues; FOD and corrosion are ambient-condition driven;hence TMF, creep, and rupture are the prime candidates for damage control and life extension ona continuous-operation basis.

TMF, creep, and rupture have similar damage patters. The simplest patter is where the damagerate (d) is geometrically proportional to a key engine operating parameter (x), sometimes called adamage driver, as shown in Figure 1. To analyze damage mechanisms more accurately, we oftenconsider additional damage drivers. Additional damage drivers reveal more complex damagepatterns as shown in Figures 2 and 3.

Paper presented at the RTO AVT Symposium on “Ageing Mechanisms and Control:Part B – Monitoring and Management of Gas Turbine Fleets for Extended Life and Reduced Costs”,

held in Manchester, UK, 8-11 October 2001, and published in RTO-MP-079(I).

(SYB) 12-2

0.8 0.82 0.84 0.86 0.88 0.9 0.92 0.94 0.96 0.98 10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Simple Damage Pattern

x

d

Figure 1: A simple damage pattern

Figure 2: A complex damage patterns

(SYB) 12-3

0.8

0.9

1

0.50.60.7

0.80.9

10

0.2

0.4

0.6

0.8

1

x

Complex Damage Pattern

y

d

Figure 3: Another damage pattern

Generally speaking, the approaches to controlling the damage and extending component life fallinto two categories:

• Active control: changing the operating procedures pertaining to mission planning orengine control, and tracking the damage concurrently1.

• Passive control: tracking damages and adjusting maintenance practices to maximize theutilization of the service life of a component.

This paper concerns with the active control approach, specifically, extending the life of hot-section components through active engine control of TMF, creep, and rupture damages. Thisapproach is called life-extending control (LEC). The LEC concept originates from damagemitigating control research for rocket engines [1-5]. It controls engine fuel flow rate by includingdamage-reduction as an active objective. The differences between a liquid-fueled rocket engineand a gas turbine engine are summarized as follows: 1) a rocket engine has a narrow operatingenvelope, its mission profile is mostly fixed; 2) a rocket engine has much shorter firingdurations; 3) a rocket engine has much longer down time for each mission cycle; 4) a rocketengine has no air breathing provision, hence, not susceptible to contamination and corrosion.

The challenge of LEC is to maintain satisfactory levels of performance and operability whilereducing component damages. To meet this challenge, LEC is preferably designed to trim thestandard engine control logic with a limited authority.

This paper describes two methodologies to reduce the life cycle cost of gas turbine engines. Thefirst methodology reduces stress rupture/creep damage to turbine blades and stators byoptimizing damage accumulation concurrently with the flight mission. This methodology is

1 By concurrent tracking of damages we mean the time from feeding damage information back to mission planningor engine control is much shorter compared with this feedback process in the passive control approach.

(SYB) 12-4

described in Section 2. The second methodology modifies the baseline control logic of an engineto reduce the TMF damage of cooled stators during acceleration. This methodology is describedin Section 3. These methodologies have also been implemented in an actual full-authority digitalelectronic control (FADEC) unit of a small gas turbine engine to demonstrate the feasibility ofLEC. A real-time, hardware-in-the-loop (HITL) simulation has also been conducted as a part ofthe feasibility demonstration. Section 4 describes the HITL simulation.

2. Stress Rupture/Creep Damage Reduction

A typical flight mission of an aircraft consists of taxi, take-off, climb, cruise, descent andlanding. In this section, we describe the reduction of rupture damage during a specific portion ofa flight mission: cruise. Since a civil airplane flies most of the time at the cruise condition,reducing engine component damages during cruise will directly increase the service life of theengine components.

Generally speaking, increasing cruise speed reduces flight time but increases the thrustrequirement. This implies higher engine speed and temperature hence high damage rate to theturbine blades and stators. Therefore, there is trade-off among flight time, fuel cost, andaccumulated component damages during cruise. An optimization to perform this trade-offamong flight time, fuel cost, and accumulated engine component damages during cruise wasformulated and is shown in the Figure 4 below.

Figure 4: A trade-off between performance and rupture/creep damage in cruise conditions

Flight Mission

A business jet is used to demonstrate this trade-off optimization concept. A typical flightmission of this type of airplane is shown in Figure 5. There are three cruise segments in theflight mission. The first cruise segment is at altitude 41000ft, the second cruise segment is ataltitude 43,000ft, and the third cruise segment is at altitude 45,000ft. The Mach number for allthree cruise segments is 0.8.

OverallValue

FlightTime

FuelCost

Damage/Life

MinimizingOverall Value

Mach Number/Cruise Speed

(SYB) 12-5

0 50 100 150 200 250 300 350 400 4500

1

2

3

4

5x 10

4

Alti

tude

(ft)

0 50 100 150 200 250 300 350 400 4500

0.2

0.4

0.6

0.8

Mac

h

Time (min)

Figure 5: A typical flight mission of the business jet

Aircraft Model

From the equations of motion of an aircraft in level flight, the required engine thrust in cruisecondition can be determined from the following two equations:

221 VSCT dρ= (1)

221 VSCmg lρ= (2)

where ρ the density of the air, S the reference area of the aircraft, dC the drag coefficient, lC the

lift coefficient, V is the cruise speed.

The relationship between dC and lC is described by the drag-polar equation:

20 l

CCC dd β+= (3)

where the zero-lift drag coefficient 0dC and the induce drag factor β are functions of Mach

number only.

The thrust T, as a function of cruise speed and mass of aircraft, can be written as

2

222

021 2

SV

gmVSCT d ρ

βρ += (4)

Cumulative Damage In Cruise

Based on the required thrust determined by Eq. (4), cumulative component damages duringcruise are determined by using the damage model. For the first cruise segment of the missionprofile (altitude 41000ft, cruise speed 0.8 Mach, cruise time 105 min), Figures 6 to Figure 8

(SYB) 12-6

show the cumulative damages for blades and stators. Figure 9 shows the total fuel consumptionas a function of cruise Mach number and initial weight with respect to a reference initial weight

0m g.

It can be seen from these figures that the cumulative component damage during cruise increasesexponentially with respect to the Mach number. Large damage reduction can be achieved withvery small sacrifice in flight time. Total fuel consumption during cruise.

0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.90

1

2

3

4

5

6

7x 10

-4

Mach number

Dam

age:

unc

oole

d bl

ade

100%m0

90%m0

Figure 6: Cumulative damage of un-cooled blade during cruise

0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.90

0.5

1

1.5

2

2.5

3

3.5

4x 10

-4

Mach number

Dam

age:

coo

led

blad

e

90%m0

100%m0

Figure 7: Cumulative damage of cooled blade during cruise

(SYB) 12-7

0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.90

0.2

0.4

0.6

0.8

1x 10

-3

Mach number

Dam

age:

unc

oole

d st

ator

100%m0

90%m0

Figure 8: Cumulative damage of un-cooled stator during cruise

0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.92000

2100

2200

2300

2400

2500

2600

2700

2800

Mach number

Fue

l con

sum

ptio

n (lb

m)

90%m0

100%m0

Figure 9: Fuel consumption during cruise

Trade-off Optimization

To demonstrate this optimization approach, a linear objective function of flight time, fuelconsumption and cumulative damage is formulated as follows:

refrefrefrefreff

f

WF

WF

D

D

D

D

D

D

t

tJ 5

_3

34

_2

23

_1

12

_1 ααααα ++++=

(5)

where

ft : Cruise time

refft _ : Cruise time at a nominal cruise Mach number

1D : Cumulative damage for uncooled blade

refD _1 : Cumulative damage for uncooled blade at a nominal cruise Mach number

2D : Cumulative damage for cooled blade

(SYB) 12-8

refD _2 :Cumulative damage for uncooled blade at a nominal cruise Mach number

3D : Cumulative damage for cooled stator

refD _3 :Cumulative damage for uncooled stator at a nominal cruise Mach number

WF : Total fuel consumption during cruise

refWF : Total fuel consumption during cruise at a nominal cruise Mach number

iα : Weighting coefficients

Assume 101 =α , 31

432 === ααα , 15 =α . For different reference cruise Mach number 0.70,

0.75, 0.80, Table 1 below lists the optimal Mach number, the damages at the optimal cruiseMach number divided by the damages at the reference cruise Mach number, and the fuelconsumption at the optimal cruise Mach number divided by the fuel consumption at thereference cruise Mach number, for three reference Mach numbers.

Note that the objective function reaches its minimum at the reference cruise Mach number forthe reference Mach number below 0.70. This is caused by the large weighting on the cruise timein the objective function. The objective function at different Mach number for the referenceMach number 0.8 is shown in Figure 10. For the Mach numbers greater than 0.75, morereduction in Cumulative damages can be achieved with small reduction in cruise speed.

Table 1: Optimization resultsRef.Mach

OptimalMach

refDD

_1

1

refDD

_2

2

refDD

_3

3refF

F

0.70 0.70 1.0 1.0 1.0 1.00.75 0.72 0.43 0.58 0.41 0.960.80 0.77 0.32 0.48 0.30 0.94

0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.8510

15

20

25

Mach number

Obj

ectiv

e fu

nctio

n va

lue

Figure 10: Objective function value at different cruise Mach number

(SYB) 12-9

3. TMF Damage Reduction

The actual engine control logic has been modified to reduce the TMF damage during engineacceleration from ground idle to maximum power. The goal is to reduce the TMF damage whilemaintaining fast engine acceleration. Several approaches to modifying engine control logic havebeen investigated including: target speed offset, control gain increase/decrease and accelerationschedule reduction. It was found from engine simulation that acceleration schedule reduction isthe most effective.

In a typical turbine engine control, engine acceleration follows an acceleration schedule;specifically, the engine speed is controlled to follow the acceleration schedule. To reduce TMFdamage, the acceleration schedule was reduced by a certain percentage once the differencebetween the controlled speed, high pressure spool speed (NH) and the target speed is less than athreshold. This is illustrated in Figure 11 below.

N2max4

Ndot_sch_max

N2C2 (rpm)

DN

Figure 11: Illustration of acceleration schedule reduction logic

For the threshold values (DN) of 800 rpm, 1000 rpm, and 1200 rpm, the reductions in TMFdamage (in percentage) and the increase of rise time of fan speed (N1) (an indicator of enginethrust) during the engine acceleration from ground idle to maximum power are shown in Tables2 to Table 4, and in Figure 12 and Figure 13 for 50% to 90% reduction of the accelerationschedule. It can be seen that the greater the reduction of TMF damage, the greater the increasein rise time. It is also found that the relationship between the TMF damage reduction andincrease in rise time is not sensitive to the threshold values. For all three cases, significantreductions in TMF damage can be achieved with only a very small increase in rise time for N1and thrust.

Table 2: TMF damage reduction for DN=800 rpm%, reduction TMF reduction (%) Extra rise time (sec)10% 13.7 0.0620% 24.5 0.1230% 35.3 0.2240% 45.6 0.3250% 49.0 0.58

(SYB) 12-10

Table 4: TMF damage reduction for DN=1000 rpm%, reduction TMF reduction (%) Extra rise time (sec)10% 14.7 0.0620% 26.4 0.1630% 37.7 0.2840% 47.5 0.4050% 54.3 0.74

Table 5: TMF damage reduction for DN=1200 rpm%, reduction TMF reduction (%) Extra rise time (sec)10% 14.7 0.0820% 27.5 0.1830% 39.2 0.3240% 49.0 0.5050% 56.9 0.86

0 10 20 30 40 50-10

0

10

20

30

40

50

60

Percentage reduction of accel schedule (%)

TM

P d

amag

e re

duct

ion

(%)

DN=800DN=1000DN=1200

Figure 12: TMF reduction vs. reduction of acceleration schedule vs. speed threshold

(SYB) 12-11

0 0.2 0.4 0.6 0.80

10

20

30

40

50

60

Increase in rising time (sec)

TM

P d

amag

e re

duct

ion

(%)

DN=800DN=1000DN=1200

Figure 13: TMF reduction vs. increase in rise time vs. speed threshold

4. Hardware-in-the-loop Simulation

The methodologies have been implemented in an actual full-authority digital electronic control(FADEC) unit of a small gas turbine engine to demonstrate the feasibility of LEC. Real-time,hardware-in-the-loop simulations have been conducted and verified the LEC concept through thetwo life extension methodologies. Figure 14 shows the simulation environment and a datascreen.

Figure 14a: HITL simulation setup Figure 14b: real-time data display

(SYB) 12-12

5. Conclusions

This paper describes two methodologies to extend the service life of hot-section components,particularly, turbine blades and stators, by reducing the damages incurred on these components.One methodology has been designed to reduce the creep damage in cruise. The othermethodology has been designed to reduce the thermo-mechanical fatigue damage in rapidtransients. These methodologies for damage reduction and life extension have been evaluated fora small commercial turbine engine for a general aviation aircraft. Evaluation was performed byhardware-in-the-loop simulations where an actual engine full-authority digital electronic control(FADEC) unit was modified with the LEC, and it interacted with an engine simulator in realtime. The results of this evaluation show that significant reductions in these damages arepromising and the design for life extension could be considered in engine control systems. Theteam anticipates to substantiate the analytical results by carefully designed experiments andengine testing in the next phase of the program.

6. Acknowledgment

The authors want to thank Dr. Ten-Hui Guo of NASA Glenn Research Center and Mr. Robert S.McCarty of Honeywell Engines and Systems for their support during the course of this three-year program. The financial support from NASA and the generous support of technicalinformation by Honeywell are essential for the success of this program.

7. References

1. Ray A, Dai, X, Wu M-K, Carpino M., Lorenzo C., “Damage-mitigating Control of aResusable Rocket Engine”, AIAA Journal of Propulsion and Power, vol. 10, pp:225-233,1994.

2. Ray A, Wu M.K., Carpino M., Lorenzo C., “Damage-mitigating Control of MechanicalSystems: “Part I:-Concept Development and Model Formulation”, Journal of DynamicSystems, Measurement, and Control, vol. 116, pp.347-447, 1994.

3. Ray A, Wu M.K., Carpino M., Lorenzo C., “Damage-mitigating Control of MechanicalSystems: “Part II:-Formulation of an Optimal Policy and Simulation”, Journal of DynamicSystems, Measurement, and Control, vol. 116, pp.448-455, 1994.

4. Dai X., Ray A., “Damage-mitigating Control of a Resusable Rocket Engine: Part I-LifePrediction of the Main Thrust Chamber Wall”, Journal of Dynamic Systems, Measurement,and Control, vol.118, pp.401-408, 1996.

5. Dai X., Ray A., “Damage-mitigating Control of a Resusable Rocket Engine: Part II-Formulation of an Optimal Policy”, Journal of Dynamic Systems, Measurement, and Control,vol.118, pp.409-415, 1996.

(SYB) 12-13

Dr. Link JawPresident and CEOLink@scientificmonitoring.com

Link Jaw founded Scientific Monitoring, Inc. (SMI) in 1993. Hehas been involved in all aspects of corporate managementincluding projects, strategies and resources. Link has over 25years of experience in engineering, software, andmanagement. Prior to starting SMI, he worked for AlliedSignalAerospace, Link Flight Simulation, and FlightSafety Simulation. Link is the inventor of five U.S. patents. He holds an M. S.degree from the University of Michigan and a Ph. D. degreefrom Stanford University. He also completed executivemanagement training at the Tuck School of BusinessAdministration of Dartmouth College.

(SYB) 12-14

Paper 12: Discussion

Question from H Pfoertner – MTU, Germany

Creep life usage is heavily dependent on temperature. Does the cruise segment, as considered in youranalysis, really contribute a significant percentage of total creep damage?

Is it necessary to change the control laws to change cruise speeds, surely that could be readily achieved bythe pilot?

Presenter’s Reply

Your comment on cycle-independence for creep and rupture damage is correct; these types of damagedepend mostly on temperature and pressure.

Of course, it is possible for the pilot to set the cruise speed but the pilot requires some instructions from aflight management or mission computer to know the required setting. To make optimisedrecommendations, such computers require inputs of usage tracking information from a controller ordiagnostic unit like the one described.